#MLOps for #Java Developers A Hands On Guide with #Kubeflow and #Quarkus
summary
Shows how Kubeflow plus Kubernetes actually runs model development, training and serving end-to-end: Kubeflow notebooks as Jupyter-as-a-service on shared GPUs, Argo-backed pipelines that retrain models automatically from S3-like storage, and KServe inference that scales models to zero via KNative. Demonstrates a 14 MB fraud dataset pipeline that produces a model file, deploys it to KServe, and a Quarkus REST client calling that endpoint in the demo — swaps LLM providers with langchain4j on the fly; demos run successfully.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:04MLOps fundamentals
- 5:00Key MLOps concepts
- 13:10Cubeflow overview
- 17:48Java in MLOps
- 19:10Cubeflow notebooks
- 21:38Cubeflow pipelines
- 25:15Model serving with KServe
- 27:00Demo Cubeflow notebooks
- 32:41Fraud detection example
- 39:05Quarkus introduction
- 43:01Quarkus and Kubernetes
- 44:29Fraud detection with Quarkus
- 47:22LangChain4j overview
- 49:11Demo multiple models
- 55:09Demo multi-model app
- 59:12Conclusion and Q&A
related talks
22mBuilding a MLOps Platform with Kubeflow on GKE
51mMachine Learning using Kubeflow and Kubernetes
15mLightning Talk: Streamlining MLOps Pipeline With Kubeflow on Arm64 Locally
30mEnd-to-End MLOps with MLflow and Kubeflow
31mKubeflow 2026 State of the Union: Orchestrating the Full ML Lifecycle
54mEconomics and Best Practices of Running AI/ML Workloads on Kubernetes
31mKubernetes and MLOps for Scalable and Reproducible Generative AI
32mKubeflow Ecosystem: What’s Next for Cloud Native AI/ML and LLMOps